h-index is practically additive in nature. One possible explanation why your h-index dropped amidst merging your publications could be due to retraction of the paper by publishers (or authors) who have originally used your paper/s as their cited reference.
In addition to the explanation by Allan Castro : An h-index of 55 means that 55 of your papers have at least 55 citations. If this number decreased to 53, two of these papers "lost" at least 2 citations. This may be due to a retraction of citing papers from ResearchGate, or due to merging of two entries for one of the same paper into one entry at ResearchGate - sometimes, there are more than one entry for a paper, and the citations are counted twice, so that the previous value of the h-index may have been overestimated. It is also possible that some of the previous citations were misidentified, and a new run of ResearchGate's algorithm corrected this. The identification of citations by ResearchGate is sometimes wrong and very often not complete, and thus the value of the indicated h-index is also not completely reliable.